Video Shot Detection & Story Board Generation Using Video Decomposition
Saumik Bhattacharya, Sumana Gupta, K. Subramanian Venkatesh · 2015
With the increasing demand of visual media, video summarization is getting more and more importance in professional content browsing, sorting and selection. It is gaining importance in surveillance to preserve storage space and to reduce user interference in analysing huge video data. Unfortunately, most of the summarization algorithms pick the frames with maximum information to summarize the video, which often make the flow of contents difficult to predict. Video shot boundary detection, on the other hand, is essential for more sophisticated applications, like video stabilization, spatio-temporal restoration, colorization etc., which demand temporal continuity within the group of frames. In this paper, we propose a novel method to detect screen shot boundaries which divides a video in groups with spatio-temporal similarities and to generate video story board using the local feature as well as the global features present in the scene, ensuring that the content of the video remains as much intuitive as possible.